DefinePK

Abstract

Parkinson's disease is a progressive neurological disorder that affects millions worldwide, causing a variety of motor and non-motor symptoms, such as the condition known as freezing of gait (FOG). The overall health of people with Parkinson's disease can be significantly enhanced by early identification and treatment of freezing of gait (FOG). Using LightGBM (Gradient Boosting Machine), a well-liked gradient boosting ensemble method that is trained using the AutoML tool and is based on decision trees, we present a prediction model for freezing of gait in this work. Leveraging a comprehensive dataset of Parkinson's disease patient’s clinical profiles, gait patterns, and demographic information, we employed feature engineering techniques to extract meaningful predictors associated with FOG. Our results demonstrate the effectiveness of the LightGBM model in accurately predicting FOG episodes in Parkinson's patients. The model evaluation shown that LightGBM offers the best results with an accuracy of 92.31% when compared to other models like Support Vector Classifier and Logistic Regression.